Commit Graph

8742 Commits

Author SHA1 Message Date
Jim Harris
3ca8b72c9e nvme_compliance: test admin commands with fuse bits set
These commands should always fail.

Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I0abfbe4a4a107957a3271692468fe69eeb823137
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9728
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2021-10-13 07:27:14 +00:00
Jacek Kalwas
a827fd7eec json: Added support for 8 bit unsigned value converter in json
Signed-off-by: Jacek Kalwas <jacek.kalwas@intel.com>
Change-Id: Id670be974f5d07f5292d338448cb0ada9510b105
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9787
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2021-10-12 23:46:49 +00:00
Shuhei Matsumoto
45dddbc891 scsi: Decide LUN format for each LUN ID
Previously we decided which LUN format is used by the macro constant
SPDK_SCSI_DEV_MAX_LUN. However, as long as we read SAM, the LUN format
can be decided per LUN ID.

Linux host SCSI driver supports 256 LUNs per SCSI device at the
maximum. So we cannot test this fix on any actual system but we
apply this fix for the potential future cases.

Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ifa6a3b66431f5e1eade326348dd99b8b9653408b
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9664
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2021-10-12 09:10:18 +00:00
Shuhei Matsumoto
5e87727596 scsi: SCSI device supports 256 LUNs at the maximum by default
Most SCSI hosts, Linux, Windows, VMware, supports 256 LUNs per
device now, and it is not easy to test even if any other non-free
OS or driver supports more than 256 LUNs.

Hence increase the macro constant SPDK_SCSI_DEV_MAX_LUN from 64 to
256. Then we do not need to expose it publicly now. So move it to
lib/scsi/scsi_internal.h.

Update the CHANGELOG together.

Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Iacde46c3854f326eebfb8befb47d41fce383b027
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9631
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2021-10-12 09:10:18 +00:00
Shuhei Matsumoto
4265fc50d9 scsi: Manage LUNs per device by not fixed size array but linked list
Change a fixed size array to a linked list to manage LUNs per SCSI
device.

Keep the linked list sorted by LUN ID because this is necessary to
efficiently find the lowest free LUN ID or check the specified LUN is free.

To avoid traversing the linked list twice, change scsi_dev_find_free_lun()
to return the LUN which comes just before where we want to insert an new LUN.

Additionally, previously spdk_scsi_dev_add_lun_ext() had not checked if
the specified LUN ID was duplicated. Fix the bug in this patch.

Add unit test cases for the function scsi_dev_find_free_lun().

These changes will enable the following patches to increase
SPDK_SCSI_DEV_MAX_LUN from 64 to 256 without consuming additional memory.

Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I7f6f070ddc680127cf86ae255055da2d1d29e4ec
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9630
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
2021-10-12 09:10:18 +00:00
GangCao
e9030f36bc trace: add an assert for the valid history
Change-Id: I4dd97351354af8ccdd0684afda06e6f2e8889839
Signed-off-by: GangCao <gang.cao@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9828
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
2021-10-12 08:52:10 +00:00
Shuhei Matsumoto
845db70ccc scsi: Report LUNs use spdk_scsi_dev_get_first/next_lun() to iterate LUNs
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ib812c1f4c180f4173c3b3d668a98e3b23ed32899
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9629
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2021-10-11 10:08:22 +00:00
Shuhei Matsumoto
27646a08d7 vhost/scsi: Use spdk_scsi_dev_get_first/next_lun() to iterate LUNs.
This is the same effort as lib/iscsi.

By using spdk_scsi_dev_get_first_lun() and spdk_scsi_dev_get_next_lun(),
remove the dependency on SPDK_SCSI_DEV_MAX_LUN from lib/vhost.

Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ib79d2c3f8e7a5027d3d2c03f8f886c588b86009b
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9611
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2021-10-11 10:08:22 +00:00
Shuhei Matsumoto
356df80468 iscsi: iSCSI connection use spdk_scsi_dev_get_first/next_lun() to iterate LUNs
Use two new public APIs spdk_scsi_dev_get_first_lun() and
spdk_scsi_dev_get_next_lun() to manage iSCSI LUNs by linked list and
to traverse all LUNs of the SCSI device.

By these changes, we can remove the dependency on the macro constant
SPDK_SCSI_DEV_MAX_LUN from lib/iscsi.

Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I3237e7f887dd761d6812ed4313f87b6e7884ea29
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9610
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2021-10-11 10:08:22 +00:00
Shuhei Matsumoto
153373a5e7 iscsi: Pass iscsi_lun directly to the callback argument of spdk_scsi_lun_open()
By this change, we will not need to traverse LUN list or tree in the
callback to hot remove.

Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ibe72fba824553d0189b9120884aa2113599a568d
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9627
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: GangCao <gang.cao@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2021-10-11 10:08:22 +00:00
Shuhei Matsumoto
594f46d7a9 iscsi: iSCSI target uses spdk_scsi_dev_get_first/next_lun() to iterate LUNs
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I99cc770c3637d79689d61f996ae40bb4be25cb00
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9626
Community-CI: Mellanox Build Bot
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: GangCao <gang.cao@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2021-10-11 10:08:22 +00:00
Shuhei Matsumoto
8215c0acef scsi: spdk_scsi_dev_add_lun() checks if lun_id is less than max by itself
This is an effort to remove the dependency on the macro constant
SPDK_SCSI_DEV_MAX_LUN from lib/iscsi.

Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ieeff4136b16cc6bfa92614248f150bc9dfe3dc74
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9609
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: GangCao <gang.cao@intel.com>
2021-10-11 10:08:22 +00:00
Shuhei Matsumoto
f61d81e47c scsi: Add spdk_scsi_dev_get_first/next_lun() to traverse all LUNs
Add two public APIs spdk_scsi_dev_get_first_lun() and
spdk_scsi_dev_get_next_lun() to remove the dependency on the macro
constant SPDK_SCSI_DEV_MAX_LUN from lib/iscsi and lib/vhost.

Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I6546697f823fe9f4fa34e1161f5c7fa912dd2d59
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9608
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: GangCao <gang.cao@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
2021-10-11 10:08:22 +00:00
Changpeng Liu
232ed2a3d3 nvmf/vfio-user: check invalid minimal qsize when creating IO SQ/CQ
Fix issue #2175.

Change-Id: I89365f2c482026c0a0313c4e27a802c0a66bebef
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9606
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2021-10-11 01:42:33 +00:00
Changpeng Liu
091aa2b681 nvmf/vfio-user: fix potential overflow for qsize/MQES/NLB
The spec treats the sizes (MQES or qsize from create/delete
IO queue command) as a 0-based value of uint16_t, but vfio-user
treats them as 1-based value, so we need to use uint32_t to
make sure the value can't overflow.  The same for NLB(number of
logical blocks).

Change-Id: I7654b7e12234525c0fce78a713dd50097e9b3d58
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9632
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2021-10-11 01:42:33 +00:00
Changpeng Liu
318a3f0a43 nvmf/vfio-user: optimize lookup_io_q() function
The function is used to check the IO SQ or CQ exists or not,
so return bool type is better and also rename it.

Change-Id: I1af2a25255b012dcf1e1952b828617ededf9b897
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9680
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Andreas Economides <andreas.economides@nutanix.com>
2021-10-11 01:42:33 +00:00
Changpeng Liu
aa9e0ac633 nvmf/vfio-user: support delete/create SQ
Previously vfio-user will hit an error when creating a deleted IO SQ.
For lookup_io_q() function, actually we should not use the queue's
address to check the queue is exist or not, as when there is a memory
removal, we will also set the queue's address to NULL and reset it again,
so here we use the queue state to indicate the queue is exist or not.

Fix issue #2174.

Change-Id: I56bdd63947c2b30a598c427e64ff2dafe1cc1d2e
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9605
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2021-10-11 01:42:33 +00:00
peluse
504bf59148 lib/idxd: only select idxd device that are on the same socket
Prior a regular round robin could result in strange performance
if an idxd device from another socket was used.

Signed-off-by: peluse <peluse@localhost.localdomain>
Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Change-Id: Id863c79067beabe73ef89d92b3fb3c436821b97a
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9367
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Reviewed-by: Monica Kenguva <monica.kenguva@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
2021-10-07 09:26:50 +00:00
John Levon
17199cdc8f add name to fd groups
For debugging purposes, take a name for identifying fds added to a group.

Signed-off-by: John Levon <john.levon@nutanix.com>
Change-Id: If1654e56e19f7fa964446ef1b9e71debf74979d1
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9731
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
2021-10-07 09:25:48 +00:00
John Levon
2d48b7dc10 lib/event: avoid cast of event_queue_run_batch
Signed-off-by: John Levon <john.levon@nutanix.com>
Change-Id: I1f23b9ab3aa5ace791a9460267606fcb03287245
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9776
Community-CI: Mellanox Build Bot
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
2021-10-07 09:25:48 +00:00
Konrad Sztyber
f243f62450 lib/trace_parser: update object statistics
Object's statistics (its index and start timestamp) are now tracked and
updated in the trace entry object.

This is the final piece of code that had to be copied from the trace
app.  The following patch will remove that code from the application and
replace it with functions from the trace_parse library.

Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I1acaefb5a516bbc59c0549b2f66a5c52368bcd2c
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9435
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Reviewed-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2021-10-06 07:30:17 +00:00
Konrad Sztyber
189b0f0920 lib/trace_parser: method for retrieving entry count
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: If5f1b1bea0d30419be46e704ddd7c2f9556b4627
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9498
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Reviewed-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2021-10-06 07:30:17 +00:00
Konrad Sztyber
ffaec5f91c lib/trace_parser: reconstruct tracepoint arguments
Tracepoint arguments are now reconstructed when retrieving next trace
entry (possible from multiple buffers).  Similarly to previous patches,
the code is directly copied from the spdk_trace app.

Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: Ie7594d381660ef31c286b5e124d2a6df3835111e
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9434
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Reviewed-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2021-10-06 07:30:17 +00:00
Konrad Sztyber
c1f6054114 nvme: asynchronous detach without shutdown notification
Detaching a controller with `no_shn_notification` flag set will follow
the regular detach path making it asynchronous too.

Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I0b9c6c23626b4cc1cfaedb3268024776a07b9195
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9003
Community-CI: Mellanox Build Bot
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
2021-10-04 15:00:35 +00:00
Konrad Sztyber
b6ecc37298 nvme: make ctrlr detach fully asynchronous
The controller detach had asynchronous API (with async/poll), but the
register operations were synchronous, so they would block on fabrics
controllers.  In this patch, they're changed to their non-blocking
counterparts, making the detach fully asynchronous.

Signed-off-by: Jim Harris <james.r.harris@intel.com>
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I74df12ab40a54f1d675639672e03755c89768bef
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8726
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
2021-10-04 15:00:35 +00:00
Changpeng Liu
b3ddd46e0d nvmf/vfio-user: return error when deleting SQ more than once
Previously we only use an assertion to address this sceanrio.

Fix issue #2173.

Change-Id: I7a6e715977218d2a3a08c48a9935880f3fe4ec63
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9604
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Reviewed-by: Andreas Economides <andreas.economides@nutanix.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
2021-09-30 08:08:05 +00:00
Changpeng Liu
742ae4ec72 nvmf/vfio-user: check SQ doorbell is valid or not before use
According to the specification, we should also post an AER
error event for this error case.

Fix #2171.

Change-Id: Ifb2343453ea5e36ce244938a939537ee6ed1c4e1
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9584
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
2021-09-30 08:08:05 +00:00
Nick Connolly
2a51824a54 lib/thread: Use PRIu64 for portability
POSIX defines PRId64/PRIu64/PRIx64 for printing 64-bit values in a
portable way. Replace a reference to %lu to remove the assumption
about the size of a long.

Signed-off-by: Nick Connolly <nick.connolly@mayadata.io>
Change-Id: I622fd43e7acf2cb93d3ba4ba9e9367e6dd064a74
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9663
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
2021-09-29 08:58:28 +00:00
Konrad Sztyber
279b7babe4 lib/trace_parser: method for iterating over entries
Added a definition of a parsed trace entry and a function allowing for
iterating over these objects.  The difference between a parsed and a
regular trace entry is that it includes more information gathered while
processing the trace file (e.g. lcore, object statistics) and provides a
contigous buffer for trace arguments.

For now, only lcore and the pointer to the actual trace entry are
filled.  Tracepoint arguments and object statistics will be added in
subsequent patches.

Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I4d5e30a7abb4860a5ba9db46f64ceae8bd14646f
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9433
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
2021-09-29 08:57:58 +00:00
Konrad Sztyber
6727cc382b lib/trace_parser: method for retrieving tsc offset
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I11fc341b4353c01cd933d6692cf97de3a3744a84
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9432
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
2021-09-29 08:57:58 +00:00
Konrad Sztyber
4ba770720a lib/trace_parser: populate entries map
The trace file is now parsed and the entries are put in a map, sorted by
their timestamps.  The code is directly copied from the spdk_trace app,
with very little modifications.

Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I7929497ffd3079b6974f5423c82a6128db2cee98
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9431
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
2021-09-29 08:57:58 +00:00
Konrad Sztyber
41ba2b30e7 lib/trace_parser: method for retrieving trace flags
It gives user access to things like the tsc rate and tracepoint
definitions.

Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: Ib50126b331faa4508174c7cb707643a3d8db6a01
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9430
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
2021-09-29 08:57:58 +00:00
Konrad Sztyber
55f64c36c6 lib/trace_parser: map/unmap the trace file
Copied code responsible for mapping/unmapping the trace file.  The
only modifications were related with tying it to the spdk_trace_parser
object.

Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: Ia575101532c612b185bd971c69157623b52b9e81
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9429
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
2021-09-29 08:57:58 +00:00
Konrad Sztyber
d919a197d6 lib/trace_parser: add trace_parser library stubs
This library will provide functions that parse traces recorded by an
SPDK application.  This includes merging traces from multiple cores,
sorting them by their timestamp and constructing trace entries spanning
across multiple buffers.  All of these tasks are currently implemented
in the spdk_trace app, so most of its code will be moved here (this is
the reason for using C++).

The motivation for extracting this code to a library is to be able to
use it from places other than the spdk_trace app, specifically the
`scripts/bpf/trace.py` script.

The main reason for creating a separate library instead of extending
libtrace is to avoid pulling in all of its dependencies.  ISA-L is the
most problematic, as we only build it as a static library, which makes
it impossible to use with dlopen (making it unusable in scripts).

Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: If101ca3425d7404abd51b0da2031358d0be44766
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9428
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
2021-09-29 08:57:58 +00:00
Changpeng Liu
f7866a563f nvmf: consolidate AER notification into one function
Change-Id: If43e92fad60eff3e3f12cac1a8b413f5c16232fb
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9633
Community-CI: Mellanox Build Bot
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2021-09-28 12:51:06 +00:00
Changpeng Liu
86ad145bc1 nvmf: add AER error event API support
We still don't support get log page with error
information LID.

Change-Id: I92db361dc956ea3ed4f6e7bdfdca763d0fea6886
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9583
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
2021-09-28 12:51:06 +00:00
Jacek Kalwas
fb335cb29c nvmf: allow to query only specified transport
Similar is already done for json-rpc bdev_get_bdevs, it might be
useful for the upper layer which has no interest in all but only
in one specified.

Signed-off-by: Jacek Kalwas <jacek.kalwas@intel.com>
Change-Id: Ie1af1cb4778edd265914bbfdc2777f66c6c76572
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9362
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2021-09-28 07:36:45 +00:00
John Levon
180d2c1d9d nvmf/vfio-user: add transport opts to RPC
Signed-off-by: John Levon <john.levon@nutanix.com>
Change-Id: Ifa8e80e0a25af7757181f480ab0405ec902a61ff
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9596
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
2021-09-27 10:56:47 +00:00
Jim Harris
974271125d nvmf/vfio_user: free request when handling invalid IO opcs
IO commands with invalid OPCs are not freeing the
associated request object after handling the response.
This would eventually result in requests on the qpair
becoming exhausted which ends up failing the controller.

Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I7c1c46265a38b31181cd5d9a98c528816ab482d3

Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9601
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
2021-09-27 10:55:34 +00:00
Ziye Yang
34c901e308 nvme/tcp: Fix tcp_req->datao calculation issue.
When data digest is enabled for a nvme tcp qpair, we can use accel_fw
to calculate the data crc32c. Then if there are multiple
c2h pdus are coming, we can use both CPU resource directly
and accel_fw framework to caculate the checksum. Then the datao value compare
will not match since we will not update "datao" in the pdu coming order.

For example, if we receive 4 pdus, named as A, B, C, D.
   offset   data_len (in bytes)
A:  0       8192
B:  8192    4096
C:  12288   8192
D:  20480   4096

For receving the pdu, we hope that we can continue exeution even if
we use the offloading engine in accel_fw. Then in this situation,
if Pdu(C) is offloaded by accel_fw. Then our logic will continue receving
PDU(D). And according to the logic in our code, this time we leverage CPU
to calculate crc32c (Because we only have one active pdu to receive data).
Then we find the expected data offset is still 12288. Because "datao" in tcp_req will
only be updated after calling nvme_tcp_c2h_data_payload_handle function. So
while we enter nvme_tcp_c2h_data_hdr_handle function, we will find the
expected datao value is not as expected compared with the data offset value
contained in Pdu(D).

So the solution is that we create a new variable "expected_datao"
in tcp_req to do the comparation because we want to comply with the tp8000 spec
and do the offset check.

We still need use "datao" to count whether we receive the whole data or not.
So we cannot reuse "datao" variable in an early way. Otherwise, we will
release tcp_req structure early and cause another bug.

PS: This bug was not found early because previously the sw path in accel_fw
directly calculated the crc32c and called the user callback. Now we use a list and the
poller to handle,  then it triggers this issue. Definitely, it will be much easier to
trigger this issue if we use real hardware engine.

Fixes #2098

Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Change-Id: I10f5938a6342028d08d90820b2c14e4260134d77
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9612
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: GangCao <gang.cao@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2021-09-27 10:53:04 +00:00
Changpeng Liu
553a6e7ad9 nvmf/vfio-user: fix race condition when free_ctrlr()
This commit fixes a race condition when calling free_ctrlr(),
nvmf_vfio_user_close_qpair->free_qp will set controller `ctrlr->qp[qid] = NULL`
finally, when calling free_ctrlr() we also need to check `ctrlr->qp[qid]`
is NULL or not, when there are multiple IO queues, we need a lock to protect
`ctrlr->qp[qid]`.  However, the call to free_qp() in free_ctrlr() is valid
only when killing SPDK target, for all other cases, e.g: VM disconnected,
the queue pairs are already freed, so here we can process these different
cases separately, and avoid extra lock.

Change-Id: I7ab71f08bf4d737843b2af42e27b1571be0b45e9
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9351
Community-CI: Mellanox Build Bot
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Thanos Makatos <thanos.makatos@nutanix.com>
2021-09-24 07:43:12 +00:00
Changpeng Liu
5fd77e32a9 nvmf/vfio-user: unmap queue pairs before spdk_mem_unregister()
Ideally, SPDK should make sure no pending I/Os in this queue
pair are using the removed memory region. Currently we just
stop the submission path and leave a TODO comment here until
we have an asynchronous way to do this.

Also use the `<=` for the boundary check.

Change-Id: I63a2189022978811dc21f92f2599f28a5191ecd7
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9352
Community-CI: Mellanox Build Bot
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2021-09-24 07:43:12 +00:00
Mao Jiang
159fa94ad8 test/nvmf/subsystem: cases for restoring ns reservation
Add rkey checking to enhance nvmf_ns_reservation_restore().

Change-Id: I6d557adcba9bf81f954c118aa09452642318bc98
Signed-off-by: Mao Jiang <maox.jiang@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9427
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2021-09-24 07:42:51 +00:00
Changpeng Liu
7fed70f1bd env/dpdk: revert 8f7d9ec "env/dpdk: Use the DPDK device count for IOMMU mapping"
This patch revert commit 8f7d9ec.  In function vtophys_iommu_init(), we
can use `dev->drvier` in RTE_DEV_FOREACH() loop to count number of devices
probed by device driver using vfio APIs, or we will count all the PCI devices
that bind to vfio-pci driver, only the probed device's IOMMU group is added
to vfio container.

The original implementation is correct to count `g_vfio.device_ref` in
vtophys_pci_device_added(), we don't need to count it in
vtophys_iommu_device_event() callback.

Fix issue #2086.

Change-Id: Ib1502a67960a49e9a2f93823cc8ceab2e8303134
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9236
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Dong Yi <dongx.yi@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2021-09-24 07:40:21 +00:00
Konrad Sztyber
a4b7f87b61 nvme: abort queued admin requests during init
Abort any queued admin requests once admin queue gets enabled. A request
can get queued if a controller is being reset and it gets submitted
while admin qpair is being reconnected.  If these requests aren't
aborted, the init process will stall, as requests don't get resubmitted
while controller is resetting and subsequent admin commands required for
the initialization would be queued too.

Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: If456a297d2d434b3cc741816cbfb13b01d37e963
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9324
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
2021-09-24 07:38:57 +00:00
Alexey Marchuk
9381d8d399 nvme: Update spdk_nvme_ctrlr_get_memory_domain
Allow to return more than one memory domain.
This change aligns bdev and nvme API and provides
more flexibility for custom transports.

Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Change-Id: Ica9b12ad8463c361be6cb62ee2c0513eec0b486d
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9546
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
2021-09-24 07:37:45 +00:00
Alexey Marchuk
ea86c035bb nvme/tcp: NVME TCP poll group statistics
Enable dump of transport stats in functional test.
Update unit tests to support the new statistics

Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Change-Id: I815aeea7d07bd33a915f19537d60611ba7101361
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8885
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
2021-09-24 07:30:06 +00:00
Alexey Marchuk
7d589976f2 nvme_tcp: Check return value of spdk_sock_group_poll
This function may return an error and we should handle it

Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Change-Id: I996ceb6e300bd9527385aded9068b2841e94a20d
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9278
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
2021-09-24 07:30:06 +00:00
Shuhei Matsumoto
c9ab7ae4b3 nvme_tcp: Use qpair->poll_group only if it is not NULL
nvme_transport_poll_group_remove() clears qpair->poll_group. Hence
we should not use it after that.

Fixes #2170

Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I57ceee8c66684e2d02b51b8a0f3d66aacbcb9915
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9560
Community-CI: Mellanox Build Bot
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: GangCao <gang.cao@intel.com>
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Dong Yi <dongx.yi@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
2021-09-24 07:29:05 +00:00
John Levon
d42557401c lib/thread: improve error message for wrong thread context
Signed-off-by: John Levon <john.levon@nutanix.com>
Change-Id: Id591f8566df1d6522742461501f750aa9a52a190
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9480
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
2021-09-24 04:45:11 +00:00